JP4801518B2 - Charged particle beam microscopic method and charged particle beam apparatus - Google Patents

Description

  The present invention relates to a charged particle beam apparatus for measuring a dimension of a sample, an inspection apparatus used for pattern inspection of a semiconductor device, and the like.

  There are SEM (Scanning Electron Microscope), STEM (Scanning Transmission Electron Microscope), TEM (Transmission Electron Microscope: Transmission Electron Microscope), and the like as devices for visualizing the sample shape with nm (nano meter) accuracy. SEM / STEM is an apparatus that forms an image by raster scanning an electron beam converged to the nm order on a sample, detecting a signal generated from an electron beam irradiation region, and synchronizing it with the raster scanning. A TEM is a device that irradiates a sample with a parallel electron beam and magnifies and projects the electrons transmitted through the sample onto a camera or a fluorescent screen using an electromagnetic lens. In recent years, with the miniaturization of semiconductor device structures, there is a need for dimensional management of several tens of nm width by high resolution SEM, dimensional management of several nm width by STEM, that is, dimensional management using medium and high magnification images and defect inspection at medium and high magnification. It has increased. In order to accurately determine the size of the sample, the shape of the defect, and the like from the images obtained by these apparatuses, an accurate magnification of the image with respect to the sample is necessary. As a highly accurate magnification calibration technique, there is a technique described in Patent Document 1. First, the magnification with respect to a sample is calibrated using a standard sample including a repetitive pattern with a known period. Next, using a sample having a fine structure, a first image obtained by actually measuring the magnification with respect to the sample is recorded, a second image with an unknown magnification with respect to the sample is recorded, and a second image with respect to the first image is recorded using image analysis. Analyze image magnification. The magnification of the second image relative to the sample is obtained from the magnification of the first image relative to the sample and the magnification of the second image relative to the first image. Thereafter, the magnification is measured in the entire magnification range by repeating the magnification analysis using the second image as the first image.

  Here, in the said patent document 1, it is assumed that the magnification in an image surface is uniform and there is no geometric distortion. However, there are actually various factors that cause the image to be geometrically distorted. Specifically, distortion caused by sample height change or retarding electric field change as described in Patent Document 2, distortion due to distortion aberration as described in Patent Document 3, deflection distortion as described in Patent Document 4, etc. There is.

  As a technique for measuring the geometric distortion as described above, Patent Document 5 uses a mark pattern having a two-dimensional periodic structure with a period of 200 nm, and scans an electron beam so as to form an angle with the mark pattern, thereby forming an interference fringe pattern. And a technique for measuring and correcting geometric distortion from the interference fringe pattern is described. As a technique for correcting geometric distortion, Patent Document 3 measures a geometric distortion caused by a sample height change or a retarding electric field change using a standard mark, and creates a correction data table based on the measured geometric distortion. A technique for correcting geometric distortion by creating and controlling electron beam scanning based on the correction data table is described. Patent Document 2 describes a technique in which distortion due to distortion of an electro-optic lens is measured with a sample having an orthogonal line with a line width of about 5 μm to 0.5 μm and corrected with a correction lens.

JP 2006-058210 A JP 2002-181336 A JP2000-040481 JP2002-251975 JP2003-022733

  In the prior art, an image of a standard sample having a known structure is taken, and the distortion caused by the apparatus is measured by measuring the distortion of the standard sample image. The measured geometric distortion error depends on the dimensional error of the known structure as well as the image processing analysis error. Here, as standard samples used in the prior art, lattice images of a low magnification standard sample having a periodic structure with a pitch of several hundred nm, a gold single crystal (0.102 nm), a mica single crystal (1.0 nm), etc. are high. There is a standard sample for magnification. These dimensional errors are all 1% or less. However, there is currently no standard sample for intermediate magnification, that is, a sample having a periodic structure pitch of several tens to several nm and a dimensional error of 1% or less, and the dimensional error of the standard sample for medium magnification is about 5%. . The geometric distortion measurement error cannot be made smaller than the dimensional error of the standard sample. Therefore, the geometric distortion cannot be measured with high precision at a magnification at which there is no highly accurate standard sample, and high-precision correction cannot be performed. In addition, when performing a defect inspection or the like of a sample, if a difference image is calculated using an image that is not sufficiently corrected, that is, an image having geometric distortion, the geometric distortion may be erroneously recognized as a defect.

  An object of the present invention is to provide a technique for measuring and correcting geometric distortion with high accuracy even at a magnification at which a highly accurate standard sample does not exist.

  In the present invention, an absolute strain measurement standard sample having a periodic structure with a known dimension and a relative strain measurement sample having a fine structure are used. First, absolute strain at the first magnification is measured using a standard sample for measuring absolute strain having a periodic structure. A relative strain measurement sample having a fine structure is photographed at a first magnification and a second magnification. An expanded / contracted image is generated by scaling the image captured at the first magnification to the second magnification. The relative distortion of the second magnification with respect to the first magnification is measured using the stretched image and the image taken at the second magnification. The absolute distortion of the second magnification is obtained from the absolute distortion of the first magnification and the relative distortion of the second magnification. Thereafter, by replacing the second magnification with the first magnification and repeating the above steps, the absolute distortion at an arbitrary magnification can be obtained. Based on the obtained measurement result, the absolute distortion at an arbitrary magnification is corrected.

  According to the present invention, it is possible to measure and correct geometric distortion with high accuracy even at a magnification at which there is no highly accurate standard sample. Therefore, the length measurement accuracy is improved without variation at an arbitrary magnification. Furthermore, since the geometric distortion can be corrected with high accuracy, it is possible to prevent misperception even during the defect inspection of the sample, and the defect detection efficiency is improved.

  In this embodiment, an example in which the geometric distortion correction technique is applied to STEM / SEM is shown. FIG. 2 shows a basic configuration diagram of the STEM / SEM used in this embodiment. An electron gun 11 that generates the primary electron beam 31 and a control circuit 11 ′ that controls the acceleration voltage and extraction voltage of the primary electron beam 31. An irradiation lens 12 that converges the primary electron beam 31 and a control that controls its current value. Circuit 12 ′, a diaphragm 13 for controlling the divergence angle of the primary electron beam 31, a control circuit 13 ′ for controlling the position of the diaphragm, an axis deviation correcting deflector 14 for controlling the incident angle with respect to the sample 30, and a current value thereof Control circuit 14 ', a stigmator 15 for correcting the beam shape of the primary electron beam 31 incident on the sample 30, and a control circuit 15' for controlling the current value thereof, the irradiation region of the primary electron beam 31 incident on the sample 30 The image shift deflector 16 for adjusting the current and the control circuit 16 'for controlling the current value thereof, the scanning deflector 17 for raster scanning the primary electron beam 31 incident on the sample 30, and the control circuit 17 for controlling the current value thereof. 'Adjusting the focal position of the primary electron beam 31 with respect to the sample 30 Objective lens 18 and control circuit 18 ′ for controlling the current value thereof, sample stage 19 for setting the position of the sample 30 and control circuit 19 ′ for controlling the position thereof, and electron detection for detecting the electron beam 32 generated from the sample 30 And an image forming circuit 22 ′ for forming a STEM / SEM image from the detected electron beam signal and raster scanning signal, and a computer 29 equipped with control software and image processing software. Each control circuit and image forming circuit are command-controlled by a computer 29. A plurality of electron beam detectors 22 are mounted on the apparatus, and among the electron beams emitted in front of the sample 30, a bright field detector 52 that detects low-angle scattered electrons and a control unit 52 ′ that controls the bright field detector 52 ′, A dark field detector 53 for detecting high-angle scattered electrons and a control unit 53 'for controlling the same, a detector 54 for detecting reflected electrons and / or secondary electrons emitted behind the sample 30, and a control unit 54' for controlling the detector 54 ' Is installed. An image formed with electrons emitted to the front of the sample 30 is an STEM image, and an image formed with electrons emitted to the rear of the sample 30 is an SEM image. Hereinafter, only the STEM image will be described for the sake of simplicity. SEM images and STEM images differ only in the types of detected electrons, and the geometric distortions contained in the images are the same.

  First, a process for obtaining an STEM image by irradiating a sample with an electron beam from an optical system using the apparatus of FIG. 2 will be described. The primary electron beam 31 is extracted from the electron gun 11 with the extraction voltage V1, and the acceleration voltage V0 is applied. A direction substantially parallel to the optical axis of the mirror body is defined as a Z direction, and a plane substantially orthogonal to the optical axis is defined as an XY plane. The thinned sample 30 is placed on the sample stage 19 and a primary electron beam 31 is incident from the Z direction. The primary electron beam 31 converged to the nm order using the irradiation lens 12 is raster scanned on the sample 30 using the scanning deflector 17. The electron beam detector 22 detects the electron beam 32 emitted from the sample 30 by the incidence of the primary electron beam 31. A STEM image is formed in synchronization with the detected electron beam signal and the raster scanning signal of the scanning deflector 17.

  FIG. 3 shows an example of the control circuit 17 ′ of the scanning deflector 17. The scanning of the electron beam is digitally controlled, and an X scanning control value and a Y scanning control value are assigned to each pixel. That is, it has an X scanning control value table and a Y scanning control value table. The waveform generation unit generates an X scanning control signal and a Y scanning control signal based on the control value table. After the X-scan control signal and Y-scan control signal are converted from digital to analog (DA), they pass through the rotation angle setting unit that changes the scanning direction of the electron beam, and the magnification setting unit that controls the scanning range of the electron beam. Each is sent. Here, the geometric distortion generated due to the waveform distortion generated in the electronic circuit and the hysteresis of the deflector is referred to as scan distortion.

  FIG. 3 also shows a control unit 16 'of the image shift deflector 16 used for geometric distortion correction described later. The image shift deflector 16 is a deflector provided for fine adjustment of the visual field position, and is designed to act on the trajectory of the incident electron beam 31 in the same manner as the scanning deflector 17. Specifically, as the configuration in which the position of the deflector and the number of turns are the same, the coil wire is double-wrapped, and one is used for the image shift and the other is used for the scanning signal. A control value corresponding to the visual field movement amount is sent from the computer 29 to the image shift deflector 16. The sent control value is D / A converted, then converted by the rotation angle setting unit, and sent to the image shift deflector 16.

  In the present embodiment, the geometric distortion is measured by dividing into two steps of an absolute distortion measurement and a relative distortion measurement according to the flow shown in FIG. Absolute distortion is the geometric distortion of an image relative to the sample, and relative distortion is the geometric distortion of another image relative to one image.

  In measuring the geometric strain, first, the absolute strain at the first magnification is measured using a standard sample having a periodic structure (step 1). Next, a sample having a fine structure is photographed at the first magnification and the second magnification (step 2). Then, an expanded / contracted image is formed by isotropically expanding / contracting the image captured at the first magnification up to the second magnification (step 3). Next, the relative distortion of the second magnification with respect to the first magnification is measured using the stretched image and the image taken at the second magnification (step 4). Then, the absolute distortion at the second magnification is obtained from the absolute distortion at the first magnification and the relative distortion at the second magnification (step 5). Thereafter, the second magnification is replaced with the first magnification and the above steps are repeated to obtain the absolute distortion at an arbitrary magnification. Then, based on the obtained measurement result, the absolute distortion at an arbitrary magnification is corrected by controlling each control unit with a computer having a role as a correction unit.

  In relative distortion measurement, an image pair having a common visual field is photographed, and deformation of the visual field due to geometric distortion is measured by pattern matching. Relative strain measurement does not require a standard sample with a periodic structure. This can be performed with a general-purpose optical axis adjustment sample. For example, a sample for SEM coated with a metal amorphous film (such as osmium) after dispersing latex particles on a glass substrate, a sample for TEM in which gold is deposited in fine particles on a carbon thin film placed on a network of Cu foil, etc. A sample capable of observing a fine structure from low magnification to high magnification is suitable. A plurality of samples can be used according to the magnification, and an amorphous sample for low magnification and an amorphous sample for high magnification may be prepared. This is because when the samples are switched, information on geometric distortion at the magnification can be taken over by photographing both samples under the same photographing conditions.

  Thus, by optimizing the measurement method, the relative distortion measurement error can be reduced to 1% or less. The dimensional error of a periodic structure with a pitch of several tens to several nanometers observable at a medium magnification is currently about 3%.

  Here, FIG. 18 shows an example of a geometric distortion measurement error when the medium magnification calibration is performed by the conventional technique and the present invention. Measurement accuracy is higher with indirect measurement that repeats relative strain measurement starting from a magnification measured with a standard sample with a dimensional error of 1% or less, rather than directly measuring medium-high magnification geometric distortion using a standard sample with low dimensional accuracy. It can be seen that

Details of the geometric distortion measurement method will be described below.
The absolute strain measurement is performed according to the flow of FIG. Here, a sample having a one-dimensional periodic structure is used. Let the periodic interval of the sample be a, and the angle formed by the X direction be β. A STEM image of this sample is taken, and the STEM image is divided into small regions. The interval between periodic structures in the STEM image is c, and the angle formed with the X direction is γ. First, the interval c and the direction γ of the periodic structure in each small region are obtained from the Fourier transform image of the STEM image. This is converted into a scanning line interval. The scanning line interval bx in the X direction is cx = c · cos γ [number of pixels] divided by ax = a · cos β [SI unit], so that the scanning line interval by in the Y direction is cy = c · sin γ [number of pixels]. ] And ay = a · sinβ [SI units]. When there is a geometric distortion, bx and by change in the STEM image. It is determined that the geometric distortion in the X direction is corrected when all bx are equal in each small region, and the geometric distortion in the Y direction is corrected when all by are equal.

  Next, an absolute distortion correction value table is created from the scanning line intervals bx, by in each small region. The scanning line interval table of the X scanning deflector is created by plotting the scanning line interval bx measured in each small region at the center position of each small region (FIG. 6 (a)). Next, the table of scanning line intervals bx is converted into a table of electron beam translation Dx. A reference point is selected from the scanning line interval table, an integrated value of changes in the scanning line interval bx from the reference point is obtained, and a parallel movement amount Dx of the electron beam incident position is calculated. Then, the table of the parallel movement amount Dx of the electron beam is converted into a table of the correction value Ix of the deflector control value. The relationship between the deflection control value change amount and the electron beam movement amount is measured in advance, the deflection control value change amount necessary to cancel the electron beam parallel movement amount due to distortion is obtained, and the absolute distortion correction value table is obtained. Record. Finally, complete the table by complementing the unplotted position values. For example, as shown in FIG. 6 (a), it is assumed that the interval between the scanning lines bx is wide near the center of the STEM image and the interval between the scanning lines bx is narrow near the periphery. From this distribution, it is determined that the incident position of the electron beam in the peripheral portion is shifted inward. In order to cancel this shift, it is necessary to shift the electron beam incident position in the periphery to the outside. That is, the values in the absolute distortion correction value table have the distribution shown in FIG. 6 (b). The absolute distortion correction value table for the Y-scanning deflector is created in the same procedure. The created absolute distortion correction value table is stored in the memory of the calculator 29 together with the magnification.

The sample used for the absolute strain measurement desirably has a dimensional error of about 1% or less because the processing accuracy of the interval a of the periodic structure directly affects the measurement accuracy of the interval b of the scanning line. In absolute strain measurement, in order to measure the scanning line spacing bx and by in two directions, two samples with a one-dimensional periodic structure are prepared as shown in FIG. Is generally used. In addition, as shown in FIG. 5 (a), bx and by can be measured with substantially the same accuracy by preparing one sample having a one-dimensional structure and setting β to 45 °. As shown in Fig. 5 (b), it is possible to create a sample having two types of periodic structures in one field of view. However, if the sample structure is complicated, the processing accuracy of the interval a of the periodic structure may be reduced. .
Furthermore, a lattice image can be used as a standard sample for absolute strain measurement. The lattice spacing is narrow, such as gold single crystal of 0.102 nm, and wide one of mica single crystal of 1.0 nm. Since there are a plurality of periodic structures in the image, it is preferable to set the direction of the periodic structures in parallel with the X direction and the Y direction as shown in FIG. 5 (b).

  Here, setting of the field diameter of the STEM image and the number of small regions divided in the absolute strain measurement will be described. This setting is performed according to the flow of FIG. The period interval c is obtained from the interval between peaks generated in the Fourier transform image. The measurement accuracy of the period interval c is improved as the peak interval is wider and the individual peak width is narrower. The peak interval is the widest when c is a two-pixel cycle. In actual measurement, the peak interval expands and contracts, so it is set slightly larger than 2 pixels. Assuming that the sample is a standard sample having a dimensional error of 1% or less, for example, μ scale (Japanese Patent No. 2454588), and assuming that the number of pixels of the TEM image is 1024 × 1024, the field diameter of the STEM image is (240 nm / 2 pixels) × It is calculated that about 110 μm, which is about 10% smaller than 1024 pixels = 123 μm, is appropriate. This corresponds to the field diameter at the lowest magnification in the STEM. Next, the number of divisions when the photographed image is divided into small areas is set. As the number of periods n included in the small region increases, the peak width becomes narrower and the distortion measurement accuracy improves. However, when the number of periods n is increased, the visual field diameter of a small region increases, and the spatial resolution of distortion measurement deteriorates. The size of the small area needs to be set in consideration of measurement accuracy and spatial resolution. A change in distortion measurement accuracy when the size of the small region is changed is measured in advance, and the size of the small region is set based on this. Under our imaging conditions, the measurement accuracy at 20 μ-scale cycles was approximately 0.1%, so we decided that it was appropriate to divide the size of the small area into 22 × 22 with 5μm × 5μm. When higher strain measurement accuracy is required, it is necessary to increase the period included in the small area. When the spatial resolution of the strain distribution is required rather than the strain measurement accuracy, it is necessary to reduce the period included in the small region.

  As another method for measuring absolute distortion, there is a method using moire fringes. The advantage of the method using moiré fringes is that a slight change in the scanning line interval can be visualized without using image processing. However, the method using moire fringes has a lower spatial resolution than the method using the Fourier transform, and may be appropriately selected according to the purpose of use.

  Next, relative strain measurement will be described. In the geometric distortion measurement flow shown in FIG. 1, the flow after step 2 is the relative distortion measurement flow. In the geometric distortion correction, it is better to set the first magnification as a reference lower, that is, to set a wide field of view. If the field of view of the image taken at the second magnification is wider, the periphery of the image at the second magnification is formed in a region that is not photographed at the first magnification. Relative strain cannot be measured at the periphery. If the geometric distortion at the center of the image can be complemented by the geometric distortion at the periphery, the first magnification may be higher than the second magnification, but the geometric distortion at the periphery needs to be measured directly The first magnification needs to be lower than the second magnification.

  Hereinafter, details of each step will be described with reference to FIG. In step 2, the relative distortion measurement sample is photographed at a first magnification for which geometric distortion has been measured and a second magnification for which geometric distortion has not been measured. It is assumed that the absolute distortion correction value table (Ix1, Iy1) at the first magnification has been created. The absolute magnification correction value table (Ix1, Iy1) may be used to photograph the first magnification, or the correction value table may be taken as zero. The same applies to photographing at the second magnification. In any case, the selected correction value table is recorded in the memory of the computer 29. Here, the first magnification was taken using the absolute strain correction value table (Ix1, Iy1) (FIG. 8 (a)), and the second magnification was taken with the correction value table set to zero (FIG. 8 (c)). ).

  In step 3, an expansion / contraction image (FIG. 8 (b)) is generated by isotropic expansion / contraction of the first magnification image (FIG. 8 (a)) to the second magnification. The ratio of the second magnification to the first magnification is K. Takita, T. Aoki, Y. Sasaki, T. Higuchi and K. Kobayashi, “High-Accuracy Subpixel Image Registration Based on Phase-Only Correlation”, IEICE Trans. Obtained by a high-precision magnification measurement method as described in Fundamentals, Vol. E86-A, No. 8, p1925-1934 (Aug. 2003). Based on the experimental results so far, the magnification measurement error at 400 × 400 pixels, magnification ratio ~ 1 is about 0.02%, and the magnification measurement error at magnification ratio ~ 1.5 is about 0.05%, which is more accurate than other measurement methods. Is shown to be high.

  In Step 4, the relative distortion of the second magnification image (FIG. 8C) with respect to the stretched image (FIG. 8B) is measured. FIG. 17 shows the procedure for measuring the relative strain in Step 4. First, the expandable image is divided into small areas, and the center position (x, y) of each small area is recorded. For example, attention is paid to a small area with a star in FIG. 8B. Assuming that the image at the second magnification (FIG. 8C) is distorted as shown by a solid line, the position of the star symbol is translated to the position shown in FIG. 8C. A small area with an asterisk in FIG. 8 (b) is used as a reference pattern, and the amount of translation (Dx, Dy) due to geometric distortion is stopped by pattern matching. Examples of pattern matching algorithms include phase-only correlation, standardized cross-correlation, and least squares. The translation amount table is created by plotting the translation amount measured in each small region (Dx, Dy) at the center of each small region (FIG. 8 (c)). In order to correct the measured geometric distortion by deflector control, the deflector control values (dIx12, dIy12) necessary for canceling the parallel displacement (Dx, Dy) due to geometric distortion are obtained. It is assumed that the relationship between the parallel movement amount of the visual field and the deflection control value change amount has been measured in advance. Based on this relationship, an X-deflection relative distortion correction table (FIG. 8 (d)) and a Y-deflection relative distortion correction table (FIG. 8 (e)) are created from the parallel movement amount table. The value of the position that was not plotted is complemented, and the relative distortion correction value table is completed. The completed relative distortion correction value table is recorded in the memory in the computer. Note that it is also possible to divide an image photographed at the second magnification instead of an image photographed at the first magnification into small regions and obtain a parallel movement amount due to geometric distortion using this as a reference pattern.

  In devices where the electron beam scanning is controlled in analog or in devices such as TEM (Transmission Electron Microscope) that capture images with a camera, geometric distortion is achieved by geometrically transforming the captured images in a computer. A corrected image can be obtained. In this case, the value of the position not plotted in the parallel movement amount table is complemented, and the parallel movement amount table is completed and recorded in the memory in the computer. Further, instead of the translation amount table itself, values converted into parameters necessary for geometric distortion correction by image processing may be recorded in the computer. Further, when a measurement object and a measurement method are specified as in a CD-SEM (Critical Dimension Scanning Electron Microscope), the length measurement result is corrected after measuring the sample shape using the undistorted image. The flow for correcting the length measurement result is more efficient than generating an undistorted image each time an image is captured. In this case, the amount of parallel movement due to geometric distortion is converted into a table of correction parameters of the length measurement result and recorded in the computer.

  In step 5, the absolute distortion at the second magnification is obtained. This time, in step 2, the first magnification was taken using the absolute distortion correction value table (Ix1, Iy1), and the second magnification was taken with the correction value table set to zero. In this case, since the image photographed at the first magnification can be assumed to be an image without absolute distortion, the relative distortion correction value table (dIx12, dIy12) measured at step 4 is used as the absolute distortion correction value table at the second magnification ( Ix2, Iy2). As another imaging condition, if the first magnification and the second magnification were shot in step 2 with the correction value table set to zero, the absolute distortion correction value table (Ix1, Iy1) at the first magnification was set to the second magnification. When the relative strain correction value table (dIx12, dIy12) measured in step 4 is added to the correction value table expanded and contracted together, an absolute strain correction value table (Ix2, Iy2) at the second magnification is obtained. As another imaging condition, in step 2, the first magnification uses the first magnification absolute distortion correction value table (Ix1, Iy1), and the second magnification uses the first magnification absolute distortion correction value table. When using the correction value table expanded / contracted according to the magnification of 2, adding the relative distortion correction value table (dIx12, dIy12) measured in step 4 to the expanded / contracted correction value table, the absolute strain correction at the second magnification It becomes a value table (Ix2, Iy2). The absolute distortion correction value table (Ix2, Iy2) at the second magnification is recorded in the memory in the computer 29 together with the second magnification.

  Thereafter, by repeating the relative distortion measurement using the second magnification as the first magnification, an absolute distortion correction value table at an arbitrary magnification can be obtained. Based on the absolute distortion correction value table created in the above procedure, the geometric distortion is corrected using the scanning deflector control unit 17 'or the image shift deflector control unit 16'.

  The measured geometric distortion is corrected by the following method. This is an apparatus for forming an image by raster scanning of an incident electron beam, and when electron beam scanning is digitally controlled, distortion can be corrected by scanning deflector control. In digital control, the control value of the X scanning deflector and the control value of the Y scanning deflector are assigned to each pixel. By adding a correction value that cancels the geometric distortion to these control values, the geometric distortion is corrected. For the geometric distortion correction, the scanning deflector control unit 17 'or the image shift deflector control unit 16' is used. FIG. 4A is a configuration diagram when correction is performed using the scanning deflector controller 17 '. An FM (Frame Memory) for storing a correction value table is provided together with a control value table stored in a DBC (Digital Beam Controller). The initial value of the correction value table is set to zero, and the STEM image is taken by controlling the deflector with the initial value of the control value table. A geometric distortion of the STEM image is obtained, a control value for canceling the geometric distortion is calculated, and recorded in the calculator 29 as a correction value table. This correction value table is called from the computer 29 and stored in the FM. The DA conversion value of the correction value table stored in FM and the DA conversion value of the control value table stored in DBC are added and sent to the scanning deflector 17 through the rotation angle setting circuit and the magnification setting circuit. There is also a configuration in which the DBC control value table is updated with the addition value of the control value table and the correction value table without providing the FM for the correction value table. The reason why the FM is installed separately is to make it easy to add as an option to the existing equipment. There is also a configuration in which correction is performed using the image shift deflector controller 16 '. A configuration diagram in that case is shown in FIG. The scanning signal generated based on the correction value table is sent to the image shift deflector 16 after being converted by the DA converter, the rotation angle setting circuit, and the magnification setting circuit. This configuration is also easy to add as an option.

Here, the phase-only correlation method used for the current parallel displacement measurement will be described with reference to FIG. Assuming two discrete images S1 (n, m) and S2 (n, m) with a positional deviation D = (D _x , D _y ), S1 (n, m) = S2 (n + D _x , m + D _y ). Let the two-dimensional discrete Fourier transform of S1 (n, m) and S2 (n, m) be S1 ′ (k, l) and S2 ′ (k, l). The Fourier transform _{F {S (n + D x} , m + D y)} = F {S (n, m)} Since the official is of _{exp (iD x · k + iD} y l), S1 '(k L) = S2 ′ (k, l) exp (iD _x · k + iD _y · l). That is, the positional shift between S1 ′ (k, l) and S2 ′ (k, l) is expressed by the phase difference exp (iD _x · k + iD _y · l) = P ′ (k, l). Since P ′ (k, l) is also a wave having a period of (D _x , D _y ), the analysis image P (n, m) obtained by inverse Fourier transform of the phase difference image P ′ (k, l) has (D _A δ-like peak occurs at the position of _x , D _y ). Rather than removing all amplitude information, the amplitude component of S1 ′ (k, l) · S2 ′ (k, l) ^* = | S1 ′ || S2 ′ | exp (iD _x · k + iD _y · l) A log or √ processing is applied to the image to calculate an image with suppressed amplitude components, and a δ-like peak occurs at the position (D _x , D _y ) of the misregistration vector even if the image is subjected to inverse Fourier transform Therefore, the positional deviation analysis may be performed on the image. Even if the phase difference image P ′ (k, l) is subjected to Fourier transform, a δ-like peak is generated in (−D _x , −D _y ), so that the Fourier transform image of the phase difference image P ′ (k, l) Misalignment analysis may be performed.

  Since it can be assumed that there is only a δ-like peak in the analysis image P (n, m), the position of the δ-like peak can be obtained with precision after the decimal point by calculating the center of gravity or by function fitting. Further, since the peaks other than the δ-like peak can be regarded as noise, the ratio of the δ-like peak intensity to the intensity of the entire analysis image P (n, m) can be regarded as the degree of coincidence between the images. It is difficult to evaluate the reliability of the displacement analysis result with the conventional displacement analysis method, and even if an incorrect amount of displacement is output because the frequency component necessary for the analysis is insufficient, the amount of displacement The analysis / calibration flow is advanced based on this. On the other hand, since the degree of coincidence is output in this misalignment analysis method, a function that automatically sets the lower limit of the degree of coincidence and automatically takes measures such as re-taking the image if the degree of coincidence is below the lower limit Is provided. More specifically, a flow is provided in which correction values are interpolated after automatically removing the values of small areas that could not be measured correctly. If the proportion of small areas determined to be unmeasurable exceeds the set value, an error message requesting image re-reading is displayed.

  In addition, the relative amount of relative distortion can be evaluated by referring to the degree of coincidence. The degree of coincidence between the stretchable image and the image at the second magnification is calculated. Since the degree of coincidence increases as the relative distortion amount decreases, it is determined that the relative distortion amount is substantially zero when the coincidence degree is a certain value or more. In this case, the relative distortion measurement in step 4 can be omitted. It can also be determined whether or not the relative distortion has been corrected by referring to the degree of coincidence. If the degree of coincidence between the corrected image obtained by correcting the distortion based on the relative distortion correction result and the stretched image is larger than the degree of coincidence between the image before correction and the stretched image, the relative distortion is corrected. Determined.

  When the relative distortion amount of the second magnification with respect to the first magnification is large and the relative distortion cannot be corrected to a target value or less from the result obtained by one measurement, the following repeated correction is performed. An image at the first magnification is taken based on the absolute distortion correction table, and an image at the second magnification is taken using the absolute distortion correction table. Create a relative distortion correction table in the steps 3 and 4. This relative distortion correction table is added to the absolute distortion correction table at the second magnification. This is the absolute distortion correction table at the second magnification. By repeating the above correction, the relative distortion of the second magnification with respect to the first magnification is reduced to a target value or less.

  If the amount of geometric distortion is expected to be large or higher analysis accuracy is required, measuring not only the amount of translation between images but also the amount of rotation and magnification improves the accuracy of measuring the amount of translation. . When the amount of geometric distortion is small, only the amount of translation is measured to shorten the calculation time. For example, when the recipe is created, the parallel movement amount, the rotation amount, and the magnification are measured, and when the rotation amount and the magnification are equal to or less than a predetermined value designated by the user, only the parallel movement amount is measured.

  Finally, a control screen used for geometric distortion correction will be described. FIG. 9 shows the basic configuration of the main window of the computer 29. The control bar includes a menu bar 41 in which the processing contents of the control software are arranged, a tool bar 42 in which the processing contents are arranged in icons, a figure window 43 for displaying STEM images, and a sub-window 44 for displaying and setting various parameters. From the menu bar or tool bar, you can select “Parameter Setting”, “Absolute Measurement”, “Relative Measurement” and “Correction Execution”. When “Parameter Setting” is selected, a window for setting geometric distortion measurement parameters is displayed (FIG. 10). In this window, a box is displayed for each magnification, and any one of absolute measurement, relative measurement, and unmeasured can be selected. When the magnification at which absolute measurement is selected is double-clicked, an absolute distortion measurement condition window shown in FIG. 11 (a) is displayed. The display box 47 displays a selection magnification for measuring the absolute distortion. In the input box 48, input the interval a of the periodic structure in the standard sample, the number of small area divisions, and the like. When the magnification at which relative measurement is selected is double-clicked, the window shown in FIG. 11 (b) is displayed. The display box 47 displays a selection magnification for measuring the relative distortion. In the input box 48, an image magnification, the number of small areas, and the like used as a reference for relative measurement are input. Double-clicking on the magnification for which unmeasured is selected displays the window shown in FIG. 11 (c). The geometric distortion is not measured at this magnification, and a correction value table is created based on the geometric distortion measured at another magnification. In the input box 48 of this window, a magnification for referring to the correction value table is input. Each window shown in FIGS. 11A to 11C is provided with a push button for displaying a measurement result. When this button is clicked after the geometric distortion measurement, a measurement result display window shown in FIG. 11 (d) is displayed, and the measurement result is displayed so as to overlap the STEM image. Measurement results include scan line spacing (bx, by), electron beam incident position translation (Dx, Dy), relative distortion correction value (dIx, dIy), absolute distortion correction value (Ix, Iy), coincidence Degrees are available. Which measurement result is displayed can be selected from the pull-down menu 46. Although not shown in FIG. 11, it is possible to display the measurement results alone as shown in FIG. 6 and FIG. It is also possible to display a numerical table instead of an arrow.

  Regarding the setting of the geometric distortion measurement parameter, several settings that are frequently used are prepared and can be selected from the pull-down menu 46 shown in FIG. For example, when the geometric distortion at each magnification is completely unknown, it is necessary to measure the geometric distortion at all magnifications. In that case, the setting shown in FIG. 10 (b) is selected. That is, geometric distortion measurement is performed at all magnifications. The absolute strain measurement is selected as the magnification at which the absolute strain can be measured with the standard sample, and the relative strain measurement is selected as the other magnification. As another case, geometric distortion can be predicted to some extent from the results of previous experiments. For example, it is assumed that the scan distortion changes when the magnification setting circuit is switched, and that if the same magnification setting circuit is used, the scan distortion hardly changes even if the magnification changes. In that case, geometric distortion correction is performed under the conditions shown in FIG. Here, it is assumed that four magnification setting circuits for the first range, the second range, the third range, and the fourth range are switched and used from the low magnification. In the first range, since geometric distortion caused by distortion is obvious, geometric distortion is measured for each magnification from the lowest magnification to create a distortion correction value table. The geometric distortion measurement is terminated when the geometric distortion caused by the distortion becomes a negligible magnification. The absolute distortion correction value table created at this time is used as the first range absolute distortion correction value table with the remaining magnifications. The absolute distortion correction value table for the second range is the relative distortion correction value measured from the absolute distortion correction value table for the first range, the image taken at the highest magnification of the first range, and the image taken at the lowest magnification of the second range. Create from a table. This absolute distortion correction value table is used at other magnifications in the second range. A third range absolute distortion correction table and a fourth range absolute distortion correction table are created in the same procedure, and geometric distortion correction is performed. In addition, geometric distortion measurement conditions can be set according to user needs. For example, as shown in FIG. 10 (c), absolute measurement can be performed using a lattice image observed at a high magnification. The set conditions can be saved as custom conditions.

  After setting the geometric distortion measurement conditions, the geometric distortion correction is executed according to the flow shown in FIG. When the periodic structure sample is moved into the STEM visual field and “Absolute measurement” is selected from the menu shown in FIG. 9, an image with a magnification at which absolute measurement is set is taken, and a correction value table is created. Next, when the microstructure sample is moved into the STEM visual field and “relative measurement” is selected from the menu shown in FIG. 9, an image with a magnification for which the relative measurement is set is taken, and a correction value table is created. Relative measurement is performed sequentially from a magnification close to the magnification at which absolute measurement was performed. After creating the correction value table for each magnification, when the “Execute correction” check button is turned on, the scanning deflector control unit 17 ′ or the image shift deflector control unit 16 ′ controls based on the correction value table. The geometric distortion is corrected.

  In order to notify the geometric distortion measurement status, a function is also provided for displaying a mark on the magnification icon during geometric distortion measurement in the window of FIG. 10 (FIG. 10 (c)). In addition, a function for displaying a mark on a magnification icon for which geometric distortion could not be measured is also provided (FIG. 10 (b)). In a high-magnification image, the fine structure included in the image is reduced, and distortion analysis may not be possible. The details of the measurement result are confirmed by opening the window shown in FIG.

  In this embodiment, a geometric distortion correction technique in TEM is shown. FIG. 13 shows a basic configuration diagram of a TEM used in this embodiment. The electron gun 11 that generates the primary electron beam 31 and the control circuit 11 ′ that controls the acceleration voltage and the extraction voltage of the primary electron beam 31; the irradiation lens 12 that adjusts the convergence condition of the primary electron beam 31; Control circuit 12 'for controlling, condenser aperture 13 for controlling the divergence angle of the primary electron beam 31, control circuit 13' for controlling the position of the capacitor diaphragm, and the incident angle of the primary electron beam 31 incident on the sample 30 Axis deviation correcting deflector 14 to be adjusted and a control circuit 14 'for controlling the current value thereof, a stigmator 15 for adjusting the beam shape of the primary electron beam 31 incident on the sample 30, and a control circuit 15 for controlling the current value thereof. 'The objective lens 18 for adjusting the focal position of the primary electron beam 31 with respect to the sample 30 and the control circuit 18 for controlling the current value thereof, the sample stage 19 for setting the position of the sample 30 in the sample chamber and the position thereof are controlled. Control circuit 19 'passing through sample 30 A projection lens 21 for projecting the transmitted electron beam 32 and a control circuit 21 ′ for controlling the current value thereof, an electron detection camera 26 for detecting the projected electron beam 32, and a control circuit 26 ′ for controlling the gain and offset thereof, TEM The computer 29 includes control software and image processing software. Each control circuit is command-controlled by a computer 29.

  First, a process for obtaining a TEM image using the apparatus of FIG. 13 will be described. A primary electron beam is extracted from the electron gun 11 with an extraction voltage V1, and an acceleration voltage V0 is applied. A direction substantially parallel to the optical axis of the mirror body is defined as a Z direction, and a plane substantially orthogonal to the optical axis is defined as an XY plane. The thinned sample 30 is placed on the sample stage 19 and a primary electron beam 31 is incident from the Z direction. Using the irradiation lens 12, the axis deviation correcting deflector 13, and the stigmator 14, the primary electron beam 31 is adjusted to be incident parallel to the sample at an incident angle parallel to the Z axis. When the primary electron beam 31 is incident on the thinned sample 30, most of the electrons are transmitted through the sample 30. The projection lens 21 is used to project the image plane of the transmission electron beam 32 onto the electron detection camera 26 to obtain a TEM image. The magnification of the TEM image is set by the excitation current of the projection lens 21.

The TEM image may be distorted due to the distortion of the projection lens. The geometric distortion of the TEM image can also be measured by the geometric distortion measurement method described in Example 1. The difference between TEM and STEM is the geometric distortion correction method. Since TEM captures images with a camera, geometric distortion cannot be corrected in the same way as STEM. One of the geometric distortion correction methods in TEM is a method of correcting with a geometric distortion correction lens. Correction in real time is possible by performing correction with the correction lens. Yet another correction method is correction by image processing. An image captured by the electron beam camera 26 is geometrically transformed using a calculator 29. In correction by image processing, even complex distortion can be corrected easily. It is better to determine which correction method to select based on the geometric distortion measurement result of the projection lens 21 and the skill level of the operator.
The other steps are almost the same as the geometric distortion correction in the first embodiment.

  In this embodiment, a geometric distortion correction technique in a wafer-compatible SEM is shown. FIG. 14 shows a basic configuration diagram of a wafer-compatible SEM used in this embodiment. Control circuit 11 'for controlling the acceleration voltage and extraction voltage of the electron gun 11 for generating the primary electron beam 31 and the primary electron beam 31 and the irradiation lens 12 for adjusting the convergence condition of the primary electron beam 31 and its current value are controlled. A control circuit 12 ′ for adjusting the incident angle of the primary electron beam 31 incident on the sample 30, a deflector 14 for correcting misalignment, a control circuit 14 ′ for controlling the current value thereof, and a primary electron beam incident on the sample 30. A stigmator 15 for adjusting the beam shape of 31 and a control circuit 15 ′ for controlling the current value thereof, an image shift deflector 16 for adjusting an irradiation region of the primary electron beam 31 incident on the sample 30, and a current value thereof are controlled. Control circuit 16 ', scanning deflector 17 for raster scanning the primary electron beam 31 incident on the sample 30, and control circuit 17' for controlling the current value of the primary electron beam 31. The focal position of the primary electron beam 31 with respect to the sample 30 is adjusted. Objective lens 18 and the control circuit that controls its current value 18 ′, a sample stage 19 for setting the position of the sample 30 in the sample chamber, and a control circuit 19 ′ for controlling the position, an E × B deflector 27 for deflecting electrons 51 emitted from the sample surface in a predetermined direction, and A control circuit 27 ′ for controlling the current value, a reflecting plate 28 on which the deflected electron beam collides, an electron detector 20 for detecting an electron beam emitted from the reflecting plate 28, and a control circuit 22 ′ for controlling the gain and offset thereof , A sample height sensor 34 using a laser beam 33, a control circuit 34 ′ for controlling the sample height sensor 34, a computer 29 equipped with SEM control software and image processing software. Each control circuit is command-controlled by a computer 29.

  First, steps required until an SEM image is obtained will be described. A primary electron beam is extracted from the electron gun 11 with an extraction voltage V1, and an acceleration voltage V0 is applied. A direction parallel to the optical axis of the mirror body is defined as a Z direction, and a plane orthogonal to the optical axis is defined as an XY plane. The sample 30 is inserted, the height of the sample 30 is obtained by a sample height sensor 34 using a laser, and the objective lens 18 is focused on the sample 30 by adjusting the Z position of the sample stage 19 or adjusting the control value of the objective lens 18. . This adjustment is a rough adjustment, and it is sufficient to have a matching accuracy that allows image analysis. Further, the measurement of the focal length may be performed by image processing. After the coarse focus adjustment, the field of view for adjusting the electron optical system is selected using the XY moving mechanism of the sample stage 19. Corrects misalignment, focus, and astigmatism in the field of view for adjusting the electron optics system. Next, the sample stage 19 is used to move to the photographing field of view, and after finely adjusting the focus of the objective lens 18 so that the image can be clearly observed, the image is captured.

  Recent SEM for wafers employs a decelerating electric field method (hereinafter referred to as a retarding method) in which the acceleration voltage on the primary side is increased to improve the spatial resolution and the negative voltage is applied to the sample side to decelerate the sample incident voltage. Has been. A negative voltage applied to the sample side is called a retarding voltage Vr. In the CD-SEM used for dimensional management of semiconductor devices, the retarding voltage is set so that the sample incident voltage approaches zero. In an electron beam wafer inspection system that detects a defect of a semiconductor device from a potential contrast, a retarding voltage Vr is adjusted according to the device in order to form a desired potential contrast. When the retarding voltage Vr is changed, the magnification of the SEM image with respect to the sample changes. When the retarding voltage has a distribution in the wafer plane, the SEM image magnification changes in the wafer plane. That is, geometric distortion occurs. This geometric distortion is measured and corrected by the method described in the first embodiment. The distribution of the retarding voltage Vr in the wafer surface depends on the sample stage shape. Therefore, in Example 3, it is necessary to update the correction value table according to the position of the sample stage.

Also, recent SEM for wafers is often customized for specific inspection and measurement. A DR-SEM (Defect Review Scanning Electron Microscope) is a device that detects a defective portion of a device from a difference image between a reference image and an input image. In this case, generation of an undistorted image becomes important. This is because in a distorted image, the pattern shape changes depending on the position in the image, and thus there is a possibility that the pattern shape change due to the image distortion is erroneously recognized as defective. On the other hand, it is not essential to generate a distortion-free image in an apparatus that measures the size of a designated location such as a CD-SEM. This is because there is a method of correcting the length measurement result after measuring the sample shape using the undistorted image. It is more efficient to correct the length measurement result than to perform distortion correction on the entire image. It can be said that the generation of the undistorted image is sufficient only when it is desired to observe the image in detail as in the recipe creation. It is better to set the target for geometric distortion correction according to the purpose of inspection / measurement.
The other steps are almost the same as the geometric distortion correction in the first embodiment.

  The geometric distortion correction technique in the present invention can be applied not only to an electron microscope but also to other charged particle beam apparatuses. For example, the present invention can be applied to Focused Ion Beam (FIB) in which an ion beam is finely focused and raster scanning is performed on a sample, and an electron beam or secondary ion generated from the sample is detected to visualize the sample structure. The present invention can also be applied to various probe microscopes such as Scanning Tunnel Microscopy (STM) and Atomic Force Microscopy (AFM).

The flowchart which shows a geometric distortion measurement process. Basic configuration diagram of STEM / SEM. FIG. 6 is a basic configuration diagram of a conventional scanning deflector control unit and a conventional image shift deflector control unit. FIG. 3 is a basic configuration diagram of (a) a scanning deflector control unit having a geometric distortion correction function and (b) an image shift deflector control unit having a geometric distortion correction function. It is the schematic of the sample used for an absolute geometric distortion correction | amendment, (a) 1-dimensional periodic structure, (b) 2-dimensional periodic structure, (c) The sample which has 2 types of 1-dimensional periodic structures is shown. It is explanatory drawing which shows the process of an absolute geometric distortion correction | amendment, (a) Distribution of the scanning line space | interval bx of a X direction, (b) Distribution of the correction value Ix of the deflector for X scanning, (c) Scanning line of a Y direction (D) shows an example of the distribution of the correction values Iy of the Y scanning deflector. The flowchart which shows the setting procedure of absolute distortion measurement conditions. It is an explanatory view showing the process of relative distortion measurement, (a) distortion measured image, (b) stretched image, (c) distortion unmeasured image, (d) correction value table of X scanning deflector, (d ) Shows an example of a correction value table for a Y-scanning deflector. The basic block diagram of a main window. It is a basic block diagram of a geometric distortion correction parameter setting window, and shows an example of (a) standard setting, (b) detailed setting, and (c) custom setting. FIG. 4 is a basic configuration diagram of a window for inputting (a) absolute strain measurement conditions, (b) relative strain measurement conditions, and (c) undistorted measurement conditions, and (d) a basic configuration diagram of a window for displaying distortion measurement results. The flowchart which shows a geometric distortion correction process. Basic configuration diagram of TEM. Basic configuration of SEM for wafers. Explanatory drawing of the parallel displacement measuring method. The flowchart which shows the preparation procedure of an absolute distortion correction value table. The flowchart which shows the preparation procedure of a relative distortion correction value table. Explanatory drawing which shows the difference in the geometric distortion measurement error between a prior art and this invention.

Explanation of symbols

11 ... Electron gun, 11 '... Electron gun control circuit, 12 ... Irradiation lens, 12' ... Irradiation lens control circuit, 13 ... Condenser diaphragm, 13 '... Condenser diaphragm control circuit, 14 ... Deflector for correcting axis deviation, 14' ... Axis deviation correction deflector control circuit, 15 ... Stigmeter, 15 '... Stigmeter control circuit, 16 ... Image shift deflector, 16' ... Image shift deflector control circuit, 17 ... Scanning deflector, 17 '... Scanning deflector control circuit, 18 ... Objective lens, 18' ... Objective lens control circuit, 19 ... Sample stage, 19 '... Sample stage control circuit, 20 ... Projection lens, 20' ... Projection lens control circuit, 21 ... Axis deviation correcting deflector, 21 '... Axis deviation correcting deflector control circuit, 22 ... Electron detector, 22' ... Electron detector control circuit, 23 ... scattering angle limiting aperture, 23 '... scattering angle limiting aperture control circuit , 24 ... Objective diaphragm, 24 '... Objective diaphragm control circuit, 25 ... Restricted field diaphragm, 25' ... Restricted field diaphragm system Circuit, 26 ... Electron beam detection camera, 26 '... Electron beam detection camera control circuit, 27 ... ExB deflector, 27' ... ExB deflector control circuit, 28 ... Reflector, 29 ... STEM control software and image Computer equipped with processing software, 30 ... sample, 31 ... primary electron beam, 32 ... electron beam, 33 ... laser beam, 34 ... sample height sensor using laser beam 33, 34 '... height sensor control circuit,
41 ... Menu bar, 42 ... Toolbar, 43 ... Figia window, 44 ... Subwindow, 45 ... Submenu, 46 ... Pull-down menu, 47 ... Display box, 48 ... Input box, 49 ... Push button, 50 ... Check button, 51 ... Electrons, 52 ... bright field detector, 53 ... dark field detector, 54 ... detector, 55 ... low angle scattered electrons, 56 ... high angle scattered electrons, 57 ... secondary electrons.

Claims (12)

A charged particle beam microscopic method of irradiating a charged particle beam to a sample, detecting a charged particle beam generated secondarily from the sample, and acquiring an image of the sample,
(A) photographing a first sample with a known structure at a first magnification, and measuring an absolute strain at the first magnification from the obtained image;
(B) photographing a field of view including a common region with the first magnification and the second magnification of the second sample;
(C) generating an expandable image obtained by isotropically expanding and contracting an image of the first magnification of the second sample obtained by photographing to the second magnification;
; (D) on the basis of the telescopic image and the pattern matching result between the previous SL second magnification images of the second sample, measuring the relative distortion for the first magnification of the second magnification When,
(E) obtaining an absolute distortion of the second magnification from the absolute distortion of the first magnification and the relative distortion of the second magnification;
(F) A step of correcting the obtained absolute distortion of the second magnification, and a charged particle beam microscopic method. In the charged particle beam microscopic method according to claim 1,
The charged particle beam microscopic method, wherein the first sample is a lattice image with a known lattice spacing. The charged particle beam microscopic method according to claim 1,
Determining the absolute strain at an arbitrary magnification by repeating the steps (b) to (d), with the second magnification as the first magnification after the absolute strain measurement;
And a step of correcting absolute distortion at each magnification. A charged particle source for generating a first charged particle beam;
A first electromagnetic field generation unit for guiding the first charged particle beam to the sample;
A sample stage for setting the position of the sample with respect to the first charged particle beam;
A second electromagnetic field generator for guiding the second charged particle beam emitted from the sample to the detector;
A detector for detecting a second charged particle beam;
An image forming unit that forms an image of the sample structure based on the detector output;
Wherein the first sample of known structure first by irradiating the charged particle beam to record an image taken by the first magnification by measuring the absolute distortion for the first sample in the first magnification,
Irradiating a second sample with the first charged particle beam and recording an image taken at the first magnification and an image taken at the second magnification;
Generating an expanded / contracted image isotropically expanded / contracted from the first magnification image of the second sample to the second magnification;
Measuring relative distortion of the second magnification image relative to the first magnification image from pattern matching between the stretchable image and the second magnification image of the second sample ;
A geometric strain measurement unit for obtaining an absolute strain at the second magnification from an absolute strain at the first magnification of the first sample and a relative strain with respect to the first magnification at the second magnification of the second sample; ,
A charged particle beam apparatus comprising: a correction unit that corrects the measured absolute distortion. The charged particle device according to claim 4 .
A charged particle beam apparatus comprising a computer having the correction unit therein. In the charged particle beam device according to claim 5 ,
The charged particle beam apparatus according to claim 1, wherein the computer has the geometric distortion measuring unit therein. The charged particle beam device according to claim 4 ,
The first electromagnetic field generator includes a deflector,
The charged particle beam apparatus, wherein the correction unit controls the deflector based on the measured absolute distortion. The charged particle beam device according to claim 4 ,
The second electromagnetic field generator includes an electromagnetic lens,
The charged particle beam apparatus, wherein the correction unit controls the electromagnetic lens based on the measured absolute distortion. The charged particle beam device according to claim 4 .
The charged particle beam apparatus, wherein the correction unit corrects a length measurement result obtained from an image formed by the image forming unit. An optical system for irradiating the sample with a charged particle beam;
A detector for detecting a secondary charged particle beam generated by irradiating the sample with the charged particle beam;
In a charged particle beam apparatus having an image forming unit that forms an image based on the output of the detector,
A memory for recording an image taken at a first magnification by irradiating the first sample of known structure with the charged particle beam, and a calculation unit for measuring an absolute strain with respect to the first sample at the first magnification; Have
Said memory, by irradiating the charged particle beam in the second sample to record the image captured by the first image was taken at a magnification second magnification, the second sample obtained by photographing A stretched image obtained by isotropically stretching the first magnification image to the second magnification,
The calculation unit is configured to measure the relative strain on the image of the first magnification in the image of the second magnification from pattern matching between the image taken by said telescopic image said second magnification,
The absolute distortion in first magnification of the first sample, the absolute distortion measured in the second magnification from the relative strain on the first magnification in the second magnification of the second sample was the measurement A charged particle beam apparatus which corrects absolute distortion. The charged particle beam device according to claim 10 ,
The optical system has a deflector for deflecting the charged particle beam,
The charged particle beam apparatus, wherein the calculation unit controls the deflector so as to correct the measured absolute distortion. The charged particle beam device according to claim 10 ,
Having an electromagnetic lens between the detector and the sample;
The charged particle beam apparatus, wherein the calculation unit controls the electromagnetic lens so as to correct the measured absolute distortion.